A Randomized, Controlled Trial of Ganaxolone in Patients With Infantile Spasms: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Randomized, Controlled Trial of Ganaxolone in Patients With Infantile Spasms (NCT00441896), a randomized phase 2 study. The primary outcome is Change From Baseline in Frequency of Spasm Clusters at Day 10.

2. Clinical Question

This question needs to be read in the context of a crossover design. Participants moved through sequences that included both ganaxolone and placebo, rather than remaining in a single parallel treatment group. The registry presents the primary endpoint results separately for the two sequences in each of five cohorts. It also reports a cohort-specific ANCOVA p-value for each sequence comparison, but does not post an adjusted treatment-effect estimate or confidence interval for those comparisons.

The record lists additional outcomes, including spasm-cluster frequency at Day 20, seizure-related measures, clinical-status assessments, and a developmental assessment. These give the study a broader set of questions than the primary Day 10 endpoint alone. The information posted does not establish how the secondary outcomes relate to the primary analysis or whether any were designated for confirmatory interpretation.

3. Trial Design

This was a Phase 2 randomized, double-blind, placebo-controlled study with quadruple masking and a crossover model. The registry reports enrollment of 57 participants and lists infantile spasms as the condition. The study is marked completed. Its start is listed as 2007-01 and its primary completion as 2008-05.

Rather than assigning participants to a single treatment for the entire study, the design used two treatment sequences. Sequence A was ganaxolone, placebo, then ganaxolone; Sequence B was placebo, ganaxolone, then placebo. Within each cohort, randomization to these sequences was in a 2:1 ratio. The registry describes ascending doses of both ganaxolone and placebo, with the dose level in each subsequent cohort based on experience from previous cohorts. Specific dose amounts and the detailed timing of sequence periods are not reported in the registry record.

The primary results are organized by five cohorts and their two sequences. The number of participants shown for each sequence varies across cohorts. The results table presents the mean change in spasm clusters per day and its standard deviation for each sequence. These are descriptive summaries of the groups as posted; they are not themselves the adjusted comparisons from the sponsor’s ANCOVA analyses.

The primary endpoint assessment was based on a 24-hour video-electroencephalography assessment at Day 10. Baseline was the Day 0 assessment before study drug infusion. The registry defines the efficacy population as randomized participants who received at least one efficacy evaluation at baseline and at either Visit 5 (Day 10) or Visit 8 (Day 20); only participants with data available at the indicated time point were analyzed. The endpoint’s results table gives the analyzed numbers for each sequence, but the record does not provide a full accounting of participants who were randomized, treated, excluded, or missing at each visit.

The record does not report the specific randomization procedure, the dose schedule, the duration of each period, or the details of any washout between treatments. It also does not explain how the crossover structure was represented in the ANCOVA model. Those details matter when assessing how well the analysis reflects the design, and they should not be inferred from the sequence labels alone.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Only the first 6 of 10 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Efficacy Population included all randomized participants who received at least 1 evaluation of efficacy at both Baseline (Day 0) and at either Visit 5 (Day 10) or Visit 8 (Day 20). Only those participants with data available at the indicated time point were analyzed.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Intent-to-treat (ITT) population included all randomized participants who received at least 1 dose of study medication

Only the first 6 of 11 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Frequency of Spasm Clusters at Day 10, measured over Baseline (Day 0) and Day 10.

The primary endpoint is change from baseline in the frequency of spasm clusters at Day 10, measured in spasm clusters per day. Spasm clusters were determined using a 24-hour video-electroencephalography assessment at Day 10. Baseline was the Day 0 assessment before study drug infusion, and the registry defines change as the post-baseline value minus the baseline value.

Because the change is calculated as follow-up minus baseline, a negative value indicates that the measured frequency at follow-up was lower than at baseline; a positive value indicates it was higher. This describes the direction of the change in the measured endpoint. It does not, by itself, identify which treatment caused the change or establish a treatment difference. The registry’s adjusted comparisons, rather than a comparison of isolated group means, are the relevant posted inferential results.

A measure of spasm-cluster frequency is used to quantify change in the recorded clinical activity over the specified assessment period. The 24-hour video-electroencephalography assessment provides the registry’s stated method for determining clusters at Day 10. The endpoint is continuous, and the record summarizes it with means and standard deviations.

The endpoint is specifically a Day 10 measure. The record separately lists change in spasm-cluster frequency at Day 20 as a secondary outcome, so the two time points should not be conflated. The posted primary results do not give a confidence interval for the adjusted treatment contrast. As a result, the size and precision of the estimated between-sequence difference cannot be read directly from the primary endpoint summaries alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is continuous: it is a mean change in spasm clusters per day. The registry reports a mean and standard deviation for each cohort-sequence group, and describes ANCOVA as the method used for the five reported comparisons. No adjusted effect estimate, confidence interval, or model specification is posted alongside the p-values. That distinction is important: group summaries describe the observed data, while ANCOVA is an inferential method intended to compare groups after accounting for model terms.

What does ANCOVA do? In general, analysis of covariance models a continuous outcome in relation to a group or treatment indicator and one or more covariates. Covariate adjustment can improve precision or account for relevant differences, depending on the model and data. The record identifies ANCOVA but does not report which covariates were included, how they were selected, or how the crossover design was represented. It is therefore not possible to reconstruct the sponsor’s model from the posted description alone.

Why does the crossover design matter? In general, a crossover study compares treatments within a participant across different periods, which can make within-participant information relevant to the comparison. A suitable analysis may need to account for the relation between repeated observations from the same person, as well as treatment sequence and period effects where applicable. The registry describes two sequences and a crossover model, but does not report whether or how those features entered the ANCOVA. The sequence-level means should not be mistaken for a complete account of the within-participant treatment contrast.

How should the means and standard deviations be read? A mean is a summary of the average observed change in a group. The standard deviation describes how dispersed the individual observations are around the group mean. In general, a large standard deviation relative to the mean indicates substantial variability in the observations, but it does not show the exact distribution or identify unusual values. Here, the standard deviations vary among the displayed groups, and the analyzed group sizes also vary. The summary statistics do not show the individual participant data or the association between measurements taken from the same participant.

What does a p-value tell us? In general, a p-value measures how incompatible the observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The registry posts five ANCOVA p-values for the cohort comparisons, but not the corresponding adjusted estimates or confidence intervals. The p-values therefore cannot show the size, direction, or precision of the adjusted differences by themselves.

Why are confidence intervals useful? In general, a confidence interval communicates uncertainty about an estimated effect in a way that preserves information about both direction and scale. A narrow interval indicates greater precision than a wide interval, under the model assumptions. Because no confidence intervals are posted for the adjusted comparisons, the range of effects compatible with those analyses cannot be assessed from the registry results. The absence of an interval is not evidence that the estimate is precise or that the treatment sequences are equivalent.

Why is there no recalculated contrast? A recalculation from summary statistics requires a defined comparison and sufficient data to reproduce that comparison. The displayed primary summaries are arranged by cohort and sequence, while the trial uses a crossover structure. The recalculation for a two-arm comparison is not estimable from the posted summaries. No recalculated treatment difference or planning sample size is therefore. In general, a planning sample size calculated from an effect is an illustration of assumptions, not an observed property of a completed trial and not a measure of the study’s observed power.

What can be said about statistical significance? The record does not state the trial’s decision threshold, adjustment procedure, or a prespecified multiplicity strategy. A reader may compare the posted p-values with a conventional threshold as context, but should not treat that comparison as a full account of the trial’s statistical plan. Multiple comparisons can increase the chance of observing at least one small p-value when many hypotheses are tested. In general, a multiplicity procedure changes how evidence is evaluated across a family of tests. Whether the sponsor used such a procedure is not reported in the registry record.

What does variability mean for interpretation? In general, over-dispersion means the outcome varies more than a model’s assumed distribution would predict. This can affect uncertainty estimates and model fit. The registry provides standard deviations but does not report diagnostics, residual distributions, or an assessment of over-dispersion. It would therefore be inappropriate to assert that over-dispersion occurred or to claim that a specific adjustment was needed. The variability visible in the group summaries is a reason to consider uncertainty, not proof of a particular model failure.

The analysis population also affects what the reported estimates describe. The registry defines the efficacy population using availability of baseline and follow-up assessments, with only those having data at the indicated time point analyzed. This is not a full description of the handling of missing outcomes. In general, analyzing only available observations can be affected if the chance of having an observed assessment is related to outcomes or treatment. The record does not report the amount, reasons, or statistical handling of missing data, so the potential impact cannot be quantified from this page.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Cohort 1: Sequence A: Ganaxolone/Placebo/Ganaxolon", "Cohort 1: Sequence B: Placebo/Ganaxolone/Placebo", "Cohort 2: Sequence A: Ganaxolone/Placebo/Ganaxolon", "Cohort 2: Sequence B: Placebo/Ganaxolone/Placebo", "Cohort 3: Sequence A: Ganaxolone/Placebo/Ganaxolon", "Cohort 3: Sequence B: Placebo/Ganaxolone/Placebo")
n   <- c(6, 3, 7, 4, 11, 3)
m   <- c(1, 2.7, -5.4, -1, -1.4, -5.3)
sp  <- c(20.06, 6.66, 10.97, 1.83, 13.69, 10.12)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Cohort 1: Sequence A: Ganaxolone/Placebo/Ganaxolon  6  1.000  20.060  -20.052  22.052
CARM  Cohort 1: Sequence B: Placebo/Ganaxolone/Placebo  3  2.700  6.660  -13.844  19.244
CARM  Cohort 2: Sequence A: Ganaxolone/Placebo/Ganaxolon  7  -5.400  10.970  -15.546  4.746
CARM  Cohort 2: Sequence B: Placebo/Ganaxolone/Placebo  4  -1.000  1.830  -3.912  1.912
CARM  Cohort 3: Sequence A: Ganaxolone/Placebo/Ganaxolon  11  -1.400  13.690  -10.597  7.797
CARM  Cohort 3: Sequence B: Placebo/Ganaxolone/Placebo  3  -5.300  10.120  -30.439  19.839

11. Analyses Reported by the Sponsor

The sponsor’s own primary analyses are the five ANCOVA comparisons reported by cohort. The registry posts p-values of 0.739, 0.537, 0.908, 0.254, and 0.666, but does not post the adjusted differences or confidence intervals for those comparisons. These results should therefore be described as the sponsor’s reported inferential analyses, not as estimates whose size and precision can be independently assessed from the registry entry.

A separate recalculated two-arm contrast is not displayed. The posted summaries are arranged by cohort and treatment sequence in a crossover design, and the data needed for a simple two-arm recalculation are not available. This means there is no recalculated estimate to compare against the sponsor’s results. It does not mean that the sponsor’s results have been independently verified or that the analyses are necessarily inconsistent with the summaries.

The registry does not flag a discrepancy between a sponsor analysis and a recalculation. In this case, however, the absence of a discrepancy flag should not be read as a numerical validation: no recalculated contrast is available for comparison. The most transparent account is to report the sponsor’s p-values as posted, explain that they do not give effect size or precision, and avoid inferring an unreported estimate from the group means.

12. Multiplicity, Missing Data and Analysis Populations

The record lists a range of outcomes beyond the Day 10 primary endpoint. Secondary outcomes include change in spasm-cluster frequency at Day 20; absence of hypsarrhythmia; change in clinical status on investigator and caregiver global assessments; spasm-free durations; seizure-free days; responders; and a Denver-II developmental assessment at Day 20. The registry reports eight posted secondary outcomes, nine outcome measures, and ten analyses. The information does not describe a statistical hierarchy among these outcomes or which analyses, if any, were considered confirmatory.

Five cohort-specific ANCOVA p-values are reported for the primary endpoint. The record does not state whether these p-values were adjusted for the multiple cohort comparisons, whether a global test was performed, or whether another multiplicity approach was used. In general, multiplicity matters because assessing several hypotheses can change the interpretation of individual p-values. Without a stated analysis plan, the posted results cannot be interpreted as having or not having a particular correction.

The record identifies an efficacy population that includes randomized participants with at least one efficacy evaluation at baseline and at either Day 10 or Day 20, and says that only participants with data at the indicated time point were analyzed. This describes an availability-based analysis population, but it does not provide the number missing at each visit, the reasons for missing observations, or the method used to handle them. It also does not report whether all randomized participants were included in other analyses.

In general, missing-data methods rely on assumptions about why observations are absent. If the chance of a missing assessment is related to treatment or to the unobserved outcome, an analysis of available cases can differ from an analysis of all randomized participants. This is a general concern, not a claim that missingness affected this study. The registry does not report enough information to evaluate that possibility or quantify its effect.

The sample-size illustration is also unavailable for the primary endpoint because a two-arm recalculation is not estimable from the posted summaries. The registry does not report the trial’s sample-size justification, any interim analyses, or a complete prespecified statistical plan. These gaps limit assessment of how the analyses were planned and how the multiplicity and missing-data issues were intended to be addressed.

13. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: convulsion (16.7%), hydrocephalus (16.7%), bradycardia (14.3%), lethargy (14.3%).

The registry’s adverse-event results are organized by cohort and sequence. For each group, the posted data distinguish participants at risk from participants affected by serious adverse events and other adverse events, and report deaths. The serious-event counts vary across the displayed groups; the table should be consulted for the group-specific figures. The registry reports zero deaths in each listed cohort-sequence group.

The serious-event terms highlighted in the record are Convulsion and Hydrocephalus, each listed as 16.7%, and Bradycardia and Lethargy, each listed as 14.3%. These are registry-listed percentages; the information does not give the corresponding term-specific counts or explain the denominator used for each percentage. They should not be treated as percentages of all enrolled participants without that denominator information.

In general, adverse-event summaries describe events recorded during a study and do not by themselves establish that an event was caused by a particular treatment. This is especially important in a crossover study with sequences that include both ganaxolone and placebo. The registry presentation does not show a treatment-period-specific safety comparison or provide details about event timing and attribution. The group-level figures should therefore be read as descriptive safety reporting, not as a causal comparison between treatments.

The groups are small, and the table reports results by cohort and sequence rather than as a single pooled comparison. No formal inferential safety analysis or confidence intervals are reported in the registry record. These data summarize the posted events; they do not establish a definitive safety profile or support individual medical decisions.

14. What the Result Does, and Does Not, Mean

The registry results show how mean spasm-cluster frequency changed from baseline to Day 10 in the displayed cohort-sequence groups, with standard deviations describing the spread of those changes. The sponsor also reports an ANCOVA p-value for each cohort-specific sequence comparison. These are the available primary endpoint results in the record.

The posted p-values do not provide adjusted effect sizes or confidence intervals. They therefore do not show how large the estimated differences were, which direction those adjusted differences favored, or how precisely they were estimated. Comparing the values with a conventional threshold can give limited context, but cannot replace an estimate with an interval or a clear statement of the analysis plan. The results do not establish that the treatment sequences are equivalent simply because the reported p-values are not below 0.05.

Nor do the cohort means alone establish a treatment effect. They are group-level summaries in a crossover study, and the record does not provide the individual-level paired data or the full model specification. The sequence-specific results may reflect multiple aspects of the study design and observations, but the registry does not give enough detail to separate those contributions.

The results also do not answer every question raised by the study. Day 20 spasm-cluster frequency and several other secondary outcomes are listed, but a detailed interpretation of those outcomes is outside the primary Day 10 endpoint summaries. The record does not state how secondary analyses were prioritized or how multiplicity was addressed.

In general, a completed trial can contribute information even when its registry results do not resolve the size or precision of a treatment contrast. Here, the posted data allow readers to examine the descriptive summaries, the five sponsor-reported p-values, and the broad safety counts. They do not support a definitive conclusion about treatment efficacy, equivalence, or causality, and they do not provide enough detail to independently reproduce the sponsor’s ANCOVA analyses.

15. Why This Trial Matters Statistically

This trial is statistically notable because it combines a crossover design, multiple cohorts, ascending doses, and a continuous primary endpoint. Each feature affects how the results should be summarized and compared. In particular, a sequence-level mean is not automatically the same quantity as a within-participant treatment contrast, and the registry does not give the model details needed to see how the crossover structure was handled.

The record also illustrates why effect estimates and uncertainty matter alongside p-values. Five p-values are posted, but the adjusted differences and confidence intervals are not. Readers can see the reported tests, yet cannot assess the magnitude and precision of the sponsor’s estimated comparisons from those tests alone. The table of means and standard deviations adds descriptive context but does not fill that gap.

The cohort-specific presentation is useful for showing that the results were not reported as one pooled contrast. At the same time, it raises questions about multiplicity and how the cohorts relate to the analysis plan. The record does not explain whether the cohort comparisons were intended as separate tests, whether their results were combined in another analysis, or whether any multiplicity method was used.

Finally, the efficacy population is tied to availability of measurements at baseline and follow-up, while details on missing observations are not posted. This makes the analysis population and its limitations part of the statistical interpretation rather than a minor reporting detail. The case demonstrates how registry results can support careful descriptive reading while leaving important questions about estimation, uncertainty, and analytic implementation unresolved.

16. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several limitations specific to this record constrain interpretation. First, the primary results are divided across five cohorts, with different analyzed group sizes in the cohort-sequence summaries. The record does not provide a single adjusted treatment-effect estimate that combines information across cohorts, nor does it explain whether such a combined analysis was performed. This prevents a straightforward reading of one overall treatment contrast.

Second, the crossover design is described through treatment sequences, but the ANCOVA model is not specified. The record does not report the covariates, treatment-period terms, sequence terms, dose terms, or approach to repeated observations. It is therefore not possible to determine from the entry alone how the analysis accounted for the design. The available means and standard deviations do not reveal within-participant correlations or paired differences.

Third, confidence intervals are absent for the reported primary comparisons. Without intervals and adjusted effect estimates, the precision and scale of the inferential results remain unclear. The p-values alone do not provide these quantities. The posted summary statistics are not sufficient for a simple recalculated two-arm contrast, so the page does not offer an independent numerical estimate.

Fourth, the record does not provide a complete account of missing data. It defines an efficacy population based on availability of efficacy evaluations but does not report the number or reasons for missing measurements at each time point, or describe imputation or sensitivity analyses. The effect of missingness cannot be assessed from the information available.

Finally, the registry does not report the sample-size justification, interim analyses, full statistical plan, or multiplicity strategy. These omissions leave uncertainty about the intended hierarchy of analyses and the handling of the several outcomes and cohort comparisons. They are limits of the registry reporting and should not be taken as evidence that a particular procedure was or was not used.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: A Randomized, Controlled Trial of Ganaxolone in Patients With Infantile Spasms — NCT00441896.
• Registry record last updated 2023-06-02; status last verified 2023-05.
• Results first posted 2023-06-02.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

20. Record Summary

The registry describes a completed Phase 2 randomized, double-blind, placebo-controlled crossover study of ganaxolone in participants with infantile spasms. Enrollment was 57. Participants were assigned in a 2:1 ratio to one of two treatment sequences, with ascending doses of ganaxolone and placebo across cohorts.

The primary endpoint was mean change from baseline in spasm clusters per day at Day 10, based on a 24-hour video-electroencephalography assessment. Results are shown by cohort and sequence, with means and standard deviations. Five cohort-specific ANCOVA p-values are posted, but adjusted effect estimates and confidence intervals are not.

The registry also lists secondary outcomes and adverse-event summaries. It does not report the full statistical plan, multiplicity approach, sample-size justification, or detailed missing-data methods. The posted information supports a descriptive account of the results, but leaves important questions about effect size, precision, and analysis implementation unresolved.